Pre-trained LMs have shown impressive performance on downstream NLP tasks, but we have yet to establish a clear understanding of their sophistication when it comes to processing, retaining, and applying information presented in their input. In this p
aper we tackle a component of this question by examining robustness of models' ability to deploy relevant context information in the face of distracting content. We present models with cloze tasks requiring use of critical context information, and introduce distracting content to test how robustly the models retain and use that critical information for prediction. We also systematically manipulate the nature of these distractors, to shed light on dynamics of models' use of contextual cues. We find that although models appear in simple contexts to make predictions based on understanding and applying relevant facts from prior context, the presence of distracting but irrelevant content has clear impact in confusing model predictions. In particular, models appear particularly susceptible to factors of semantic similarity and word position. The findings are consistent with the conclusion that LM predictions are driven in large part by superficial contextual cues, rather than by robust representations of context meaning.
We address the problem of enhancing model robustness through regularization. Specifically, we focus on methods that regularize the model posterior difference between clean and noisy inputs. Theoretically, we provide a connection of two recent methods
, Jacobian Regularization and Virtual Adversarial Training, under this framework. Additionally, we generalize the posterior differential regularization to the family of f-divergences and characterize the overall framework in terms of the Jacobian matrix. Empirically, we compare those regularizations and standard BERT training on a diverse set of tasks to provide a comprehensive profile of their effect on model generalization. For both fully supervised and semi-supervised settings, we show that regularizing the posterior difference with f-divergence can result in well-improved model robustness. In particular, with a proper f-divergence, a BERT-base model can achieve comparable generalization as its BERT-large counterpart for in-domain, adversarial and domain shift scenarios, indicating the great potential of the proposed framework for enhancing NLP model robustness.
This work aims to find approximate relationships to calculate the static durability of ball valve disc valves, based on material resistance and flexibility theory. Three groups were obtained, and these groups represent three mathematical charts, each
of which corresponds to specific relationships obtained to calculate the stresses generated in the dish.
the scientific researches go
towards applying practical and scientific experiments to achieve these goals.
From this point, we try in our research to create strength and increase it for
ceramic tiles at phases with lower temperature degree than what is used in
industry.
Bleaching process of cotton fabrics consider an important and
sensitive operations in wet chemical treatments of cotton fabrics,
mercerization is not less important than them and that in cases in which
is made necessary, but from the problems of t
hese processes are long
time and consumption of energy, but the ultrasound energy contribute
to reduce these problems.
Lubricant type and concentration play a very important role on metalic powder
characterestics. This research presents the outcomes of an experimental investigation
on the effect of lubricant type and content to the green compressibility, and streng
th
of green pieces produced by adding different amounts (0.5,1,1.5)% and different types
(zinc stearate,graphit,poley phinil alcohol) of lubricant to the iron powder, and mixing
them for one hour in V shape blender ,then pressing them to different densities
(75,80,85)% out of the the theoretica density.
The aim of this research to study the Abrasion on carbonate rocks which comes from specific mines in Syria, and possibility finding correlation between Abrasion test(L.A) and some physical properties of carbonate rocks .
Results or Finding:
- Most
of studied carbonate rocks are resistance(18-41%).
- There is correlations between some physical properties of carbonate rocks and Abrasion test(L.A) .
Recently worldwide researches have been devoted to the use of steel fibers recovered from used tires in concrete. In Syria the amount of recovered steel from used tires is estimated about 6000 tons/year.
For this purpose a bead wire having a diamete
r of 0.8 mm from burnt tires was extracted and used. Three mixes with cement content 300- 350 -400kg/m3 were produced incorporating three different volumes of fiber 0.5%, 1.0%, and 1.5% and three different length 30-40-60 mm. The concrete obtained by adding these fibers evidenced a satisfactory improvement of the fragile matrix mostly in terms of toughness and post cracking behavior On the other hand it was improvements in compressive strength by steel fiber inclusion the interesting results confirm the promising application of concrete reinforced with steel fibers extracted from used tires in aircraft pavement hydraulic structures and ground slab in fabrics.
Influence of light represented by ultraviolet radiation having two
different wavelengths, ٢٥٤ nm and ٤٠٠ nm, on some Iraqi cellulose extracted
from reeds as a raw material has been studied. It was found that the natural
degree of polymerization of
cellulose fibers was affected by UV radiation. This
was estimated by measuring the relative viscosity of cellulose solution and
compared to that of cellulose fibers before treatment. This is also true for the
effect of heat on cellulose fibers especially in the presence of oxygen due to the
oxidation process.
These results were compared with the results obtained for imported
cellulose, which was showed that it contains high percent of water vapor,
whereas the unbleached cellulose had the higher percent of volatile materials.
This reflects the relation between the type of cellulose and it’s sources.
